During WW2 the US government has no problem employing leading experts in math, information theory, computability (as much as that was a thing then) and electrical engineering. The Navy ran an organization out of DC with the excellent name of Communications Supplementary Activity, Washington, or CSAW, that was home to most of the US efforts related to engineering and research of code-breaking computers and general purpose computers. There are a few funny stories here that we can relate to, in one way or another.
One is the involvement of von Neumann in the creation of what we call today the von Neumann architecture. There are two guys that we largely erase from computing history and those are the guys who go on to found Univac: Eckert and Mauchly. During WW2, Eckert invents a form of electromechanical memory that uses mercury to store information. Eckert and Mauchly realize that they can turn this into a general purpose computer where instructions and data are stored in this mercury memory and the computer works by retrieving both from the memory and acting on it with equipment that already exists. The reason we don’t call it the Eckert architecture is because one afternoon coming back from a PI meeting, Eckert and Mauchly’s government COR (a Navy LT) is killing some time in a NJ train station with von Neumann and brags about this fancy new general purpose computer they’re making. von Neumann is all over that shit and gets himself inserted into the project, where he goes to every meeting, asks systematizing questions, takes notes, and then writes a sole-author paper where he describes what their group has created. The Navy LT circulates this far and wide, Eckert and Mauchly resign in protest due to how the patents fall out between them, von Neumann, and UPenn, and the rest is history.
As the war is ending, staff at CSAW realize that they work for the government and peacetime efforts are going to be full of a lot of bad coffee, stupid meetings, and politics, and absent any Nazi ass-kicking. They all make plans to retire to civilian life where they can make fat stacks selling vacuum cleaners and inventing vending machines. Two mid level managers at CSAW realize that life will kind of suck for the US if all the expertise on how to make code breaking computers vanishes, and they wonder what to do about it.
What if, they wonder, we start a company with a generic sounding name and have the Navy award us a bunch of super classified sole source contracts? Absolutely everyone thinks this is sketchy as hell, for every reason from secrecy and security to the perceived/actual ethical conundrum of awarding sole source work to a company that is a few minutes old. However, this decision is easier to make because the big companies involved like IBM and NCR are walking away from government code breaking work because the volume is too low and absent a war, they don’t really have any interest in doing production runs in units of hundreds. Thus, Engineering Research Associates (ERA) is formed, a backroom deal with an aeronautics company that made gliders is formed, a sole source award is given to the glider company, all of the work is subbed out to ERA, and the government CORs that make the award happen retire and go to work for ERA. This story is older than the Roman empire.
If you need this to get even more familiar, read on from this excellent direct quote:
By the summer of 1946 ERA had hired 40 ex-CSAW technical staffers, and by the time the new employees arrived in St. Paul, NCML personnel were already occupying their new Minnehaha Avenue facilities. ERA also recruited engineers and physicists directly in the Minneapolis and St. Paul area. The Twin Cities area recruits, many of whom were University of Minnesota engineering graduates returning home from the war, saw newspaper ads calling for mechanical engineers, electrical engineers, and physicists. The ads did not mention the nature of the work, and they listed no phone number. The Minnehaha Avenue address, to which recruits were to come in person, and the names of the two companies were the only identifying information in the advertisements.
In the employment interviews, the ERA interviewers told recruits only that they would be employed in classified government work. The company then placed new hires in training courses while the Navy processed their security clearances. Even after starting work, most new people did not know the end purpose of the equipment they were building, for the Navy managers broke up all work into small security-compartmented work tasks, and only a few individuals at the top had a complete overview of what the new organization was doing.
This story is set in the good old days, so paired with the civilian ERA company is a military organization, the Naval Computing Machine Laboratory (NCML). NCML is an evaluator of ERA technology produced under contract for the Navy and unlike the evaluators of today NCML has operational relevance, engineering skill, and contractual teeth (today it seems that we have to pick two of those on a good day). They evaluate ERA systems and designs so that the systems can do things like be maintained in the field, be stripped into component parts to be transported, and fit through hatches and doors in ships. The requirements that NCML places on ERA were probably described by ERA staff as “insane” however the result was ERA equipment was solid and field serviceable while other computing equipment was much more brittle.
Of course you know how this story ends, too: ERA gets bought by a giant defense contractor (Remington Rand) and is kept separate until that isn’t convenient any more and now it’s only described in obscure text books.